In the 2012 elections for the House of Representatives, Republics won 54% of the races despite receiving fewer votes nationwide than the Democrats. In 2016, the Republicans eked out a narrow victory in the nationwide popular vote, 51% to 49%, but ended up controlling 55% of the seats.
In short, our representative democracy is failing at being representative. If our House of Representatives lived up to its name, the Republicans should only hold a nine-seat majority instead of 40 seats.
The main culprit in this problem is the phenomenon of gerrymandering, where the party that controls the state legislature gets to draw district lines, including those for their Congressional districts, to favor their candidates and disadvantage the opposition party. The trick is to group as many of the opposition party’s voters into as few districts as possible, while your own voters dominate the remaining districts.
Gerrymandering lets politicians handpick the voters who get to vote for them, which is exactly the opposite of democracy.
As I write this, court cases to determine the constitutionality of gerrymandering are making their way toward the Supreme Court, but even if the court strikes down the practice, I don’t believe that will be the end of the problem. It may force state legislatures to be less blatant about gerrymandering, but well-paid political consultants will earn their money finding ways around any court decision we might see.
I am dubious about most of the proposed solutions for gerrymandering, all of which seem like band-aids that don’t address the heart of the problem, a political election system that is vulnerable to gerrymandering in the first place. Non-partisan commissions seem like a good idea, but if they always work like they do in California, they tend to produce a sort of bi-partisan gerrymandering, where both parties have their own safe districts. If both parties, not just one, get to handpick their voters, that isn’t much of a step in the right direction.
There are also proposals that would take human hands completely out of the equation, literally. These computer algorithms draw maps that divide up the voters as evenly as possible, but often without concern for geography and the normal boundaries of cities and counties. A small corner of one county could find itself lumped into the same district with a town in another county on the other side of a mountain range.
As I noted in my post on third parties, the system by which we decide elections, known as “First Past the Post,” leaves us stuck with a two-party system, which leaves us stuck with gerrymandering. The best possible solution would be to find a replacement for “First Past the Post” that is, as much as possible, immune to two-party rule and gerrymandering.
Fortunately, these do exist.
The first is the simplest but probably the least satisfactory, and that is a straight up proportional system. In this system, citizens don’t vote directly for a candidate but indicate their party preference. If a party gets, for example, 45% of the votes, they will receive as close as possible to 45% of the seats in the legislature. The great advantage of this is that it eliminates the need for strategic voting, where a Green Party supporter, for instance, might feel the need to vote for a Democrat to keep the Republican candidate from winning, which is one way that “First Past the Post” enforces two-party rule.
In a proportional system, there is no disincentive against voting your conscience, but exactly the opposite. The more votes a minor party receives, the more representatives it can seat. And since there are no district lines to draw, gerrymandering is impossible.
For all these advantages, there are two disadvantages to a proportional system. The first is that there is no such thing as local representation. If you’re a Green Party voter in the middle of Nebraska, you could find that you haven’t elected any representatives who have ever been to Nebraska and don’t fully appreciate the local problems.
The second problem may have already occurred to you. If the voters only select their preferred party, who chooses the people who get to occupy the seats each party wins? The party is who. Before the election, each party publishes a list of their candidates, in their order of preference. The candidate at the top of the list gets the first seat, and so on down the line. If you are number forty-six on the list and your party receives forty-five seats, then you don’t get seated.
Since the parties control who’s on these lists, that gives the parties’ national committees tremendous power to reward “team players” and punish insurgents and mavericks. There are a couple of upsides. The insurgents would be free to break off and form their own party. Also, the selection of candidates and their order could be put to a vote by party members, but I don’t see the national committees of the major parties giving up that much power and control.
The second alternative to “First Past the Post” has certain disadvantages compared to a proportional system. It isn’t quite as friendly to minor parties. It isn’t automatically as immune to gerrymandering as a proportional election would be. Most of all, compared to proportional and “First Past the Post,” it’s way more complicated.
The system is known as Single Transferable Vote (STV) and it’s one of the “ranked” voting systems you might have heard Governor Howard Dean talking about. In a ranked voting system, a voter doesn’t just pick a single candidate for an office but can rank multiple candidates in order of preference. The other difference between STV and what we currently have is that STV features multi-member districts. Rather than five smaller districts each sending a single representative, you would have a larger district sending five representatives.
The primary advantage is that, not only is this system far less susceptible to gerrymandering if implemented correctly, but it will produce a more representative result from each district. If, for example, you had five heavily Republican districts in a red state like Alabama under the current system, each district would send a Republican to Congress, leaving the Democrats in those states with no presentation. Under STV, those five districts combined into the one would probably still elect three or four Republicans, but Republican domination would never reach 100% even in the reddest districts, allowing the Democrats in that district to elect at least one of their own. The opposite would be true, of course, in heavily Democratic areas like Los Angeles, the Bay Area, Chicago, and New York. Republicans from those districts would finally have representation in Congress.
Spread to the entire country, this would produce a Congress that represents the political demographics of the nation as closely as possible.
From the voter’s perspective, it’s simply a matter of selecting your favored candidates and ranking them in order of preference. How those votes transfer into election results is where this method is complicated.
First, an STV election is not a race to 50%, or even to see who gets the “most” votes. To win, you must reach a threshold based on the number of votes cast and the number of seats up for election. If you have five seats, the threshold is the number of votes divided by six. If it was three seats, you’d divide by four. There is a mathematical reason for using one more than the number of seats, but the practical reason is that it is more likely to produce a truly representative result than if we simply used the number of seats.
First, everyone’s number one choices are counted. In most elections, two or three candidates will likely exceed the threshold on their first-choice ballots. What happens then is where STV can throw people for a loop. Since this is not a race to see who gets the most votes, the excess first-choice ballots would be disenfranchised if we didn’t consider these voters’ second choices. So, when a candidate exceeds the threshold, the excess ballots are redistributed according to their second choices. If the second choice has already surpassed the threshold, then third or fourth choices can be considered.
Once all the excess votes are redistributed, we start eliminating candidates from the bottom up, taking the candidate with the least votes and redistributing their ballots according to second and third choices (fourth and fifth, if necessary). Repeat this process until we have filled all available seats with candidates who have passed the threshold. If we find ourselves with one seat not filled and no more candidates who have met the threshold, the final seat is awarded to high vote-getter who didn’t reach the threshold.
Got all that?
Like I said, it is more complicated than simply awarding the election to the person with the most votes, but complexity is the price we would have to pay for a truly representative government.
You might have noticed that I used five seats in most of my examples and there is a good reason for this. Electing five representatives per district is the minimum size for this system to gerrymander-proof. It’s also the maximum size if we don’t want our typical election to have a ridiculous number of candidates.
Why? Reasons. Math reasons which I’m not good enough at math to explain.
It isn’t necessary, however, for all districts to have the same number of seats, so long as each seat represents approximately the same number of people. A district with three seats representing 300,000 people would be fine so long as a district with five seats represented 500,000 people. This way, we could have sparsely populated areas with three representatives rather than have ridiculously vast districts just to justify sending five representatives.
Conversely, in a densely populated urban area, rather than carve up the city into small blocks, we could have districts with seven or nine seats, again as long as each seat represented the same number of people as each of three seats in a rural district. This would be a rare exception, hopefully, because a district with more seats means elections with more candidates. Having nine candidates would likely mean as many as twenty candidates up for election, which is probably as much as we could ask voters to keep track of.
What about minor party candidates? It does seem likely that this system would elect as many minor party candidates as a straight proportional system. That’s probably true, but STV would elect many more minor candidates than “First Past the Post” does now, even if it only elected a handful. I will also admit that, because larger urban districts with seven or nine seats are more likely to elect fringe parties than a typical district of five seats or a rural district of three seats, left-wing minor parties might do better under this system than right-wing minor parties, although I suspect that right-wing fringe parties might do surprisingly well in sparsely-populated rural districts. With five seats up for grabs in the typical district, it’s always possible that the major parties will run out of candidates, leaving room for a Green Party candidate or a Libertarian to slip through.
I would be negligent if I didn’t point out any system that nurtures the development of minor parties means that parties representing extreme points of view might actually gain a few seats in the House. The necessary price of greater democracy could include a white nationalist or other alt-right party gaining seats alongside a candidate representing a Black Lives Matter party, for example.
As long as we are on the subject of ranked voting systems, they don’t just work for multi-seat elections like STV but could also serve to make the popular election of the President of the United States more like to produce the best possible result.
The system for a presidential election, known as the Alternative Vote or Instant Runoff, is basically STV for a single seat. When the first-choice ballots are counted, if one candidate has 50% of the votes cast, the election is over, and that candidate wins. If not, we start eliminating the bottom dwellers and redistribute their ballots according to second and third choices until one candidate appears on at least 50% of the ballots cast.
In a presidential election (or gubernatorial or senatorial) election, the Alternative Vote eliminates the need to worry if a vote for one’s favored candidate might end up helping an opposing candidate. The Spoiler Effect, the fatal flaw of “First Past the Post,” is a thing of the past. One could vote for a Jill Stein or a Bernie Sanders without having to worry that you might help elect a Donald Trump.
The other advantage, and it’s important, is that the Alternative Vote guarantees that the winner appears on a majority of the ballots. Even if the winner was everyone’s first choice, it will be a candidate that that 50% + 1 of the electorate is okay with.
The ranked voting system approved by Maine voters in the 2016 election, but which may not survive a court challenge, badly missed the point of such a system, because it implemented the system for both the primaries and the general election. In a ranked system, because it removes the spoiler effect, allows multiple candidates from the same party to run in the same election without penalty. This effectively removes any need for primary elections. The only reason to have a nomination process in the first place is that, under “First Past the Post,” multiple candidates from one party (or even one side of the political spectrum) would be political suicide.
Picture, if you will, the U.S. Presidential election process without the Electoral College and with the Alternative Vote. Because the Alternative Vote eliminates the problem of the spoiler effect, there would be no problem even if we had all of the candidates who ran in the 2016 primaries, four Democrats and seventeen Republicans, running in the general election. Imagine if there wasn’t a tortuously long build-up to the primaries, and then the primaries. What if the entire election, from declarations of candidacy to the final vote, was just a few months, maybe only a few weeks, long? Not only would the electorate be far less tired of politics by election day, but the election itself would cost a lot less (assuming we haven’t done something about the corrupting influence of money in politics).
One logical question to ask about adopting Single Transferable Vote for the house is how to divide up 435 seats when states like Wyoming only have one representative. We wouldn’t have to. 435 is not constitutionally magic number but was fixed at that number by an act of Congress when Arizona and New Mexico joined the Union in 1912. Changing over to STV would obviously require a constitutional amendment, so increasing the number of representatives would simply be a part of that process.
The reality is that, even with the current system, the number 435 is now badly out of date given the current population of the various states. Being both a political nerd and a math nerd, I did some rough calculations and I figure that somewhere around 620 seats in the House would allow for better representation. If we switched to STV, the number would have to be considerably higher. Very crude calculations, using three representatives from a single district in the smallest state (Wyoming), puts the number somewhere around 1,600, with roughly 200 coming from California alone. I’ll admit that this probably won’t go over well in the smallest, reddest states in the Union.
Of course, that leads to another obvious question: if we have to significantly increase the number of representatives in the state, where are all these people going to sit? I’m not sure how many more than 435 the House chamber could comfortably hold. However, if you follow my previous advice and abolish the Senate, then the House has a whole other chamber to expand into.
Is implementing STV in the United States a realistic idea? Probably not. It would require a constitutional amendment and most of the states that would need to ratify the amendment would be surrendering considerable political power to a handful of the most populous states. Is it a good idea? Without a doubt. Any other solution to gerrymandering amounts to little more than a band-aid.